arazzo: 1.0.1
info:
  title: Sensible Upload URL Extract And Poll
  summary: Generate a Sensible-signed upload URL for a document type, then poll the extraction id until results are ready.
  description: >-
    The Sensible-hosted upload variant of asynchronous extraction. The workflow
    asks Sensible for a presigned upload_url scoped to a document type, returns
    that URL and the extraction id, and then polls the Retrieve extraction by ID
    endpoint until the status is COMPLETE. The actual PUT of the document bytes
    to the returned upload_url happens out of band against Amazon S3 and is not
    a Sensible API operation, so it is documented as an input expectation rather
    than modeled as a step. Every step spells out its request inline, including
    the Bearer authorization.
  version: 1.0.0
sourceDescriptions:
- name: extractionsApi
  url: ../openapi/sensible-extractions-api-openapi.yml
  type: openapi
workflows:
- workflowId: upload-url-extract-and-poll
  summary: Generate a Sensible upload URL for a document type and poll the resulting extraction to completion.
  description: >-
    Requests a presigned upload_url for the supplied document type and config,
    surfaces the upload_url and extraction id for an out-of-band PUT, then polls
    the extraction id until Sensible reports COMPLETE.
  inputs:
    type: object
    required:
    - apiKey
    - documentType
    - configName
    properties:
      apiKey:
        type: string
        description: Sensible API key used as the Bearer token.
      documentType:
        type: string
        description: The document type to extract from.
      configName:
        type: string
        description: The config to use for extraction.
      contentType:
        type: string
        description: Content type of the document you will PUT to the upload_url (e.g. application/pdf).
        default: application/pdf
      documentName:
        type: string
        description: Optional filename echoed back in the extraction response.
  steps:
  - stepId: generateUploadUrl
    description: >-
      Request a Sensible-signed upload_url for the supplied document type and
      config, returning the extraction id used to retrieve results.
    operationId: generate-an-upload-url-with-config
    parameters:
    - name: Authorization
      in: header
      value: "Bearer $inputs.apiKey"
    - name: document_type
      in: path
      value: $inputs.documentType
    - name: config_name
      in: path
      value: $inputs.configName
    requestBody:
      contentType: application/json
      payload:
        content_type: $inputs.contentType
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      extractionId: $response.body#/id
      uploadUrl: $response.body#/upload_url
      status: $response.body#/status
  - stepId: pollStatus
    description: >-
      After the document is PUT to the upload_url out of band, poll the
      extraction by id until Sensible reports the COMPLETE status.
    operationId: retrieving-results
    parameters:
    - name: Authorization
      in: header
      value: "Bearer $inputs.apiKey"
    - name: id
      in: path
      value: $steps.generateUploadUrl.outputs.extractionId
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      status: $response.body#/status
      parsedDocument: $response.body#/parsed_document
      coverage: $response.body#/coverage
    onSuccess:
    - name: extractionComplete
      type: end
      criteria:
      - context: $response.body
        condition: $.status == "COMPLETE"
        type: jsonpath
    - name: keepPolling
      type: goto
      stepId: pollStatus
      criteria:
      - context: $response.body
        condition: $.status == "WAITING" || $.status == "PROCESSING"
        type: jsonpath
  outputs:
    extractionId: $steps.generateUploadUrl.outputs.extractionId
    uploadUrl: $steps.generateUploadUrl.outputs.uploadUrl
    status: $steps.pollStatus.outputs.status
    parsedDocument: $steps.pollStatus.outputs.parsedDocument